The Basenji Club of America African Stock Project

Project Library


OBSERVATIONS OF BASENJI COAT COLOR

by Leon C. Standifer

  In Liberia we could distinguish about three general strains of Basenjis. Within these strains there was still more variation, and our group of interested owners were constantly involved in a selective breeding programs. At the same time, we all attempted to study the genetics of various traits. This was extremely difficult because we were continually adding native lines in which we had to rely entirely on phenotype.
  The only studied character which appeared to be simple and not quantitative was coat color. We felt that it was controlled by a single pair of genes with black dominant to red. In our Liberian breedings, a black x red cross usually gave a 3:1 ratio of black to red. (To my knowledge no tri-colors were ever produced and we did not see or hear of any in the country.) This would be a classic Mendelian ratio of single gene complete dominance. After returning to the United States and discussing the problem with Geneticists, I began to doubt our concept of complete dominance; and the recent red Fula line crosses which produced blacks supported the doubts. Modern Geneticists are concluding that very few characters follow the simple Mendelian ratios, so I would like to suggest the following possible means of coat color inheritance.
  It seems probable that the Basenji coat color must be produced by a multi-gene complex with black lacking complete dominance (i.e., having incomplete dominance). As an arbitrary case, suppose that the color is due to three gene pairs which we will label A, B, and C. Further suppose that to be black a Basenji must carry one dominant in each gene pair. The other component of the pair could be either dominant or recessive without affecting the color. Thus Aa Bb Cc, AA BB CC, Aa BB Cc and a few other combinations would all give black phenotypes; all that is needed is one dominant in each gene pair.
  In this type of inheritance pattern, then, all other gene combinations would be red: Aa Bb cc, AA bb CC, aa bb cc, etc. As long as one gene pair lack a dominant, the result is a red coat color.
  Then theoretically, a red AA Bb cc crossed with a black AA BB CC would yield black progeny in a 3:1 ratio of black to red. This would explain our observations in Liberia which we interpreted as simple complete dominance. The two matings of red American Basenjis into our Kiki line have followed this ratio.
  Using the same scheme, as red Aa Bb cc crossed with a red aa bb Cc could produce black offspring (genotype Aa Bb Cc). The ratios from such a cross would 7 red to 1 black. Possibly this is how the blacks were produced in the Fula line crosses.
  Again using the same scheme, it would be possible to cross a red AA BB cc and a red Aa Bb cc and get no blacks at all. Perhaps this explains why no blacks occurred before the Kiki and Fula lines were introduced.
  All the above postulations are probably over simplified. For instance, there may be 4 or more gene pairs involved instead of 3 and this would considerably reduce all of the above frequencies. These schemes are offered for comments and criticisms by persons with more experience in Basenji breeding.
  In Liberia we also attempted to study the genetics of the bark, but learned little except that it seems quantitative. We avoided dogs which had an "appreciable bark," but since no one knew where the line was drawn between a "woof" and a "bark", decisions were difficult. I now realize that our standards on this point were not as critical as those of Basenji breeders here.
  Our selections for bone structure and general conformation were rather good, partly because these characters were easily recognized. All of the Liberian lines are long in the loin, but not having other lines for comparison, most of us did not realize this.
  The curl of the tail was always a problem; the tails in the Liberian strains seldom curled as tightly as we would liked. As I mentioned before, the tail is of little importance to the native who sometimes docks it altogether. An interesting commentary on this fact can be seen on a Liberian postage stamp issued for the 1960 Olympic Games. Pictured on the stamp is a typical native hunter with his typical country dog - a beautiful Basenji, but a nearly straight tail.
  In comparing the Liberian Basenjis with the Basenji in America, there is a noticeable difference in temperament. The Liberian dog seems more relaxed, docile and easily trained. Oddly enough, this may be related to a legal characteristic of the tribal Liberian. Most natives enjoy legal "palaver" and will sue their neighbor at the slightest provocation. A dog which chases the neighbors chickens or snaps at the children can cost the owner several months wages. So here again the native system of selection comes into play - they eat the dogs that won't mind. Perhaps the true reason for this difference in temperament is more complex. Lawsuits, as a common practise, seem to have reached the tribal areas only in the last generation; although they may have been developing along with the nationalistic movement for about 60 years. This may not have been enough time to develop such a trait in the dogs; it may only have been factor in maintaining it.
  I should point out that none of these remarks is strongly significant genetically. They represent observations of possibly 50 or 60 specimens at the most and seldom from repeated matings. To establish validity, we should have a minimum of 100 individuals from the same mating. Of course this is a virtual impossibility with Basenjis, requiring about 20 repeated matings, but the validity could be increased by observations from 3 or 4 repeated matings.



The Official Bulletin of the Basenji Club of America
Nov.-Dec. 1964 pp. 9-10
Copyright © 1964 BCOA, All Rights Reserved